Mailman 3 September 2009 - NumPy-Discussion

glumpy: fast OpenGL numpy visualization + matplotlib integration
by Nicolas Rougier Sept. 28, 2009

Sept. 28, 2009

Hi all, glumpy is a fast OpenGL visualization tool for numpy arrays coded on top of pyglet (http://www.pyglet.org/). The package contains many demos showing basic usage as well as integration with matplotlib. As a reference, the animation script available from matplotlib distribution runs at around 500 fps using glumpy instead of 30 fps on my machine. Package/screenshots/explanations at: http://www.loria.fr/~rougier/coding/glumpy.html (it does not require installation so you can run demos from within the glumpy directory). Nicolas

3 3

unpacking bytes directly in numpy
by Thomas Robitaille Sept. 28, 2009

Sept. 28, 2009

Hi, To convert some bytes to e.g. a 32-bit int, I can do bytes = f.read(4) i = struct.unpack('>i', bytes)[0] and the convert it to np.int32 with i = np.int32(i) However, is there a more direct way of directly transforming bytes into a np.int32 type without the intermediate 'struct.unpack' step? Thanks for any help, Tom

3 2

Error building docs
by Michael Droettboom Sept. 28, 2009

Sept. 28, 2009

Anybody know why I might be seeing this? Cheers, Mike [~/builds/numpy_clean/doc] > make html mkdir -p build touch build/generate-stamp mkdir -p build/html build/doctrees LANG=C sphinx-build -b html -d build/doctrees source build/html Running Sphinx v1.0 (hg) 1.4.dev7413 1.4.0.dev7413 /home/mdroe/usr/lib/python2.5/site-packages/Sphinx-1.0dev_20090915-py2.5.egg/sphinx/application.py:166: FutureWarning: A plot_directive module is also available under matplotlib.sphinxext; expect this numpydoc.plot_directive module to be deprecated after relevant features have been integrated there. mod = __import__(extension, None, None, ['setup']) loading pickled environment... not found [autosummary] generating autosummary for: reference/arrays.classes.rst, reference/arrays.dtypes.rst, reference/arrays.indexing.rst, reference/arrays.interface.rst, reference/arrays.ndarray.rst, reference/arrays.rst, reference/arrays.scalars.rst, reference/c-api.array.rst, reference/c-api.config.rst, reference/c-api.coremath.rst, reference/c-api.dtype.rst, reference/c-api.generalized-ufuncs.rst, reference/c-api.rst, reference/c-api.types-and-structures.rst, reference/c-api.ufunc.rst, reference/distutils.rst, reference/index.rst, reference/internals.code-explanations.rst, reference/internals.rst, reference/maskedarray.baseclass.rst, reference/maskedarray.generic.rst, reference/maskedarray.rst, reference/routines.array-creation.rst, reference/routines.array-manipulation.rst, reference/routines.bitwise.rst, reference/routines.ctypeslib.rst, reference/routines.dtype.rst, reference/routines.dual.rst, reference/routines.emath.rst, reference/routines.err.rst, reference/routines.fft.rst, reference/routines.financial.rst, reference/routines.functional.rst, reference/routines.help.rst, reference/routines.indexing.rst, reference/routines.io.rst, reference/routines.linalg.rst, reference/routines.logic.rst, reference/routines.ma.rst, reference/routines.math.rst, reference/routines.matlib.rst, reference/routines.numarray.rst, reference/routines.oldnumeric.rst, reference/routines.other.rst, reference/routines.poly.rst, reference/routines.random.rst, reference/routines.rst, reference/routines.set.rst, reference/routines.sort.rst, reference/routines.statistics.rst, reference/routines.testing.rst, reference/routines.window.rst, reference/ufuncs.rst Failed to import 'numpy.matlib': no module named numpy.matlib WARNING: [autosummary] failed to import 'numpy.__array_priority__': no module named numpy.__array_priority__ WARNING: [autosummary] failed to import 'numpy.acorrelate': no module named numpy.acorrelate WARNING: [autosummary] failed to import 'numpy.distutils.misc_util.get_numarray_include_dirs': no module named numpy.distutils.misc_util.get_numarray_include_dirs WARNING: [autosummary] failed to import 'numpy.generic.__squeeze__': no module named numpy.generic.__squeeze__ building [html]: targets for 76 source files that are out of date updating environment: 996 added, 0 changed, 0 removed Exception occurred:[ 0%] reference/arrays.classeslass File "/home/mdroe/usr/lib/python2.5/site-packages/docutils/nodes.py", line 471, in __getitem__ return self.attributes[key] KeyError: 'numbered' The full traceback has been saved in /tmp/sphinx-err-QYFjBP.log, if you want to report the issue to the author. Please also report this if it was a user error, so that a better error message can be provided next time. Send reports to sphinx-dev(a)googlegroups.com. Thanks! make: *** [html] Error 1 -- Michael Droettboom Science Software Branch Operations and Engineering Division Space Telescope Science Institute Operated by AURA for NASA

2 1

python reduce vs numpy reduce for outer product
by Erik Tollerud Sept. 26, 2009

Sept. 26, 2009

I'm encountering behavior that I think makes sense, but I'm not sure if there's some numpy function I'm unaware of that might speed up this operation. I have a (potentially very long) sequence of vectors, but for examples' sake, I'll stick with three: [A,B,C] with lengths na,nb, and nc. To get the result I want, I first reshape them to (na,1,1) , (1,nb,1) and (1,1,nc) and do: >>>reduce(np.multiply,[A,B,C]) and the result is what I want... The curious thing is that >>>np.prod.reduce([A,B,C]) throws ValueError: setting an array element with a sequence. Presumably this is because np.prod.reduce is trying to operate elemnt-wise without broadcasting. But is there a way to make the ufunc broadcast faster than doing the python-level reduce? (I tried np.prod(broadcast_arrays([A,B,C]),axis=0), but that seemed slower, presumably because it needs to allocate the full array for all three instead of just once). Or, if there's a better way to just start with the first 3 1d vectorsand jump straight to the broadcast product (basically, an outer product over arbitrary number of dimensions...)?

2 3

np.any and np.all short-circuiting
by Citi, Luca Sept. 26, 2009

Sept. 26, 2009

Hello I noticed that python's "any" can be faster than numpy's "any" (and the similarly for "all"). Then I wondered why. I realized that numpy implements "any" as logical_or.reduce (and "all" as logical_and.reduce). This means that numpy cannot take advantage of short-circuiting. Looking at the timings confirmed my suspects. I think python fetches one element at the time from the array and as soon as any of them is true it returns true. Instead, numpy goes on until the end of the array even if the very first element is already true. Looking at the code I think I found a way to fix it. I don't see a reason why it should not work. It seems to work. But you never know. I wanted to run the test suite. I am unable to run the test on the svn version, neither from .../build/lib... nor form a different folder using sys.path.insert(0, '.../build/lib...'). In the first case I get "NameError: name 'numeric' is not defined" while in the second case zero tests are successfully performed :-) What is the correct way of running the tests (without installing the development version in the system)? Is there some expert of the inner numpy core able to tell whether the approach is correct and won't break something? I opened a ticket for this: http://projects.scipy.org/numpy/ticket/1237 Best, Luca In the following table any(x) is python's version, np.any(x) is numpy's, while *np.any(x)* is mine. '1.4.0.dev7417' x = np.zeros(100000, dtype=bool) x[i] = True %timeit any(x) %timeit np.any(x) x = np.ones(100000, dtype=bool) x[i] = False %timeit all(x) %timeit np.all(x) ANY i any(x) np.any(x) *np.any(x)* // 6.84 ms 831 µs 189 µs 50000 3.41 ms 832 µs 98 µs 10000 683 µs 831 µs 24.7 µs 1000 68.9 µs 859 µs 8.41 µs 100 7.92 µs 888 µs 6.9 µs 10 1.42 µs 832 µs 6.68 µs 0 712 ns 831 µs 6.65 µs ALL i all(x) np.all(x) *np.all(x)* // 6.65 ms 676 µs 300 µs 50000 3.32 ms 677 µs 154 µs 10000 666 µs 676 µs 36.4 µs 1000 67.9 µs 686 µs 9.86 µs 100 7.53 µs 677 µs 7.26 µs 10 1.39 µs 676 µs 7.06 µs 0 716 ns 678 µs 6.96 µs

4 9

Tuple outer product?
by Mads Ipsen Sept. 26, 2009

Sept. 26, 2009

Is there a numpy operation on two arrays, say [1,2,3] and [4,5,6], that will yield: [[(1,4),(1,5),(1,6)],[(2,4),(2,5),(2,6)],[(3,4),(3,5),(3,6)]] Any suggestions are most welcome. Mads -- +------------------------------------------------------------+ | Mads Ipsen, Scientific developer | +------------------------------+-----------------------------+ | QuantumWise A/S | phone: +45-29716388 | | Nørresøgade 27A | www: www.quantumwise.com | | DK-1370 Copenhagen, Denmark | email: mpi(a)quantumwise.com | +------------------------------+-----------------------------+

6 18

Getting number of *characters* in dtype='U' array
by Michael Droettboom Sept. 26, 2009

Sept. 26, 2009

Is there a way to get the number of characters in a fixed-size 'U' array? I can, of course, parse dtype.str, or divide dtype.itemsize by the size of a unicode character, but neither seems terribly elegant or future proof. Does numpy provide (to Python) a method for getting this that I'm just missing? In [7]: x = np.array(u'1234') In [8]: x.dtype Out[8]: dtype('<U4') In [9]: x.dtype.str Out[9]: '<U4' In [10]: x.dtype.itemsize Out[10]: 16 Cheers, Mike -- Michael Droettboom Science Software Branch Operations and Engineering Division Space Telescope Science Institute Operated by AURA for NASA

3 4

genfromtxt to structured array
by Timmie Sept. 25, 2009

Sept. 25, 2009

Hello, this may be a easier question. I want to load data into an structured array with getting the names from the column header (names=True). The data looks like: ;month;day;hour;value 1995;1;1;01;0 but loading only works only if changed to: year;month;day;hour;value 1995;1;1;01;0 How do I read in the original data? Thanks, Timmie

2 1

Re: [Numpy-discussion] F2PY error : ... on platform 'posix' with 'gcc' compiler
by Arthur Bousquet Sept. 25, 2009

Sept. 25, 2009

Thanks a lot for you help. Now it is working, I downloaded the gfortran from http://hpc.sourceforge.net/. Best regards, Arthur On Fri, Sep 25, 2009 at 3:49 PM, Arthur Bousquet <arth.bousquet(a)gmail.com>wrote: > How can I add the flags ? > > Thank > > Arthur > > > On Fri, Sep 25, 2009 at 3:46 PM, Robert Kern <robert.kern(a)gmail.com>wrote: > >> On Fri, Sep 25, 2009 at 14:44, Arthur Bousquet <arthbous(a)indiana.edu> >> wrote: >> > $ file libsw.so gives : >> > >> > libsw.so: Mach-O bundle i386 >> > >> > How can I change it ? Is it because of my gfortran ? >> >> Your gfortran is probably still configured to make x86 executables by >> default. Try recompiling libsw.so with the flags >> >> -arch x86 -arch x86_64 >> >> -- >> Robert Kern >> >> "I have come to believe that the whole world is an enigma, a harmless >> enigma that is made terrible by our own mad attempt to interpret it as >> though it had an underlying truth." >> -- Umberto Eco >> _______________________________________________ >> NumPy-Discussion mailing list >> NumPy-Discussion(a)scipy.org >> http://mail.scipy.org/mailman/listinfo/numpy-discussion >> > >

2 1

F2PY error : ... on platform 'posix' with 'gcc' compiler
by Arthur Bousquet Sept. 25, 2009

Sept. 25, 2009

Hello, I have a Mac with snow leopard 10.6. I compiled Python2.6.2, numpy1.3 and matplotlib from the sources. I tried to compiled a library (named libsw), that I have made in fortran, with f2py but I got the following error : running build_ext error: don't know how to compile C/C++ code on platform 'posix' with 'gcc' compiler make[1]: *** [libsw] Error 1 make: *** [libsw] Error 2 Noteworthy, my library is correct because I can compile it with f2py on my Ubuntu desktop. Does anyone knows where this error comes from ? Thank you. Best regards, Arthur

2 21